Marie-Hélène Budworth

For the past few years, I have attended a conference at Wharton at the University of Pennsylvania each spring.  It is always interesting and energizing. Given that it is a practitioner oriented conference, it gives me a chance to hear from leaders in various fields.  This year’s keynotes included Mary Bara, CEO of GM, Stewart Butterfield, co-founder and CEO of Slack (formerly of Flickr), and Howie Roseman, VP of the Philadelphia Eagles.  And while the talks from these giants are clearly interesting and insightful, there are always a few ‘stars’ that quietly inspire.  This year was no different.

The theme of the conference is data.  What can you do with data?  And every once in a while, the answer is that you can do something magical.

A clear stand out for me this year was the work being done at Crisis Text Line.  In a short, 15 minute talk, Bob Filbin reminded me of the power and potential of data analytics.  Bob and his team at Crisis Text Line have spent the last five years dedicated to helping and saving the lives of young people in trouble.  In 2013, Nancy Lublin made the simple observation that young people prefer to communicate using texts, and decided to build a crisis intervention service that spoke to young people on this platform.  Bob went on to describe how the, using data from historical texts, they were able to develop algorithms that allowed them to quickly identify and respond to individuals at dire risk.  There were simple yet power findings.  For example, the presence of ‘tylenol’ or ‘advil’ in the initiating text almost always signalled an active suicide event.  Today, the service saves 20 people’s lives EVERY day.  The algorithm allows them to detect high risk cases with 86% accuracy in just 39 seconds.  I am in awe.

This year’s conference was a wonderful reminder that not all data is about the bottom line, general performance improvement or the like.  Understanding patterns in human behaviour can support amazing and impressive outcomes.  As I plan projects for the coming year, I will keep this example in mind.